I continued watching the “Nanopore Sequencing Ultra Rich Data for Cancer Research” webinar. It featured Sayonika Mohanta, the Market Segment Manager for Methylation with Oxford Nanopore Technologies (ONT). Mathilde Fiser from the Curie Institute presentation had the title “Transforming Cancer Care: Redefining Cancer Characterization and Predisposition Insights through Nanopore Sequencing.” Fiser explained that the identification of germline alterations has been accelerated by SNV detection with multigene panels. With Nanopore sequencing and adaptive sampling, a BED file can be loaded containing the region of interest to sequence. This procedure results in target enrichment without amplification. The team developed a workflow. They begin with 2 ug of genomic DNA, fragment with G-Tubes from Covaris, and then use the SQK-LSK 114 ligation-based sequencing kit. They divide the library into two parts: load once and run for 24 hours. After this time, they wash and do a nuclease treatment for 2 hours. Then, they load with the remaining library and run again. The data is then analyzed using the bioinformatics tool they developed: NanoCliD. The bioinformatics pipeline uses several variant calling programs. Fiser presented results from a patient with a heterogenous germline duplication of BRACA1 exons 18-20 confirmed by MLPA. Given that RNA analysis would take some time, the team used their adaptive sampling approach. The BED file contained BRACA1 + 120 genes + 5 kb flanking. Two days after launching sequencing, the team had the results. The mean genomic depth was 2.77X, and their mean coverage on targeted regions was 22.55X. Sanger sequencing confirmed the breakpoints. The entire process took ten days. Fiser then described a DNA methylation study. Hypermethylation of MLH1 results in silencing and instability. Two patients were studied using adaptive sampling. They created a BED file with MLH1 and took 10 kb on each side and several other related genes. This approach helped them analyze the methylation levels in these two patients. Next, Fiser spoke about phasing. They applied this approach to a patient to determine which allele(s) the variants were found. They used a similar approach and created a BED file to keep 1.6% of the genome. The team sequenced and used NanoCliD to analyze the phased MUTYH variants. Fiser then spoke about their pediatric brain tumor analyses. The team created a panel of genes for adaptive sampling. They selected genes based on literature and analyses. Each gene had a 5 kb flanking sequence, and the panel accounted for 1.6% of the genome. The clinical example described by Fiser used a frozen brain biopsy sample. Within one week, the team was able to generate results. In a second case, they could also simultaneously detect methylation and fusions. Fiser then spoke about medulloblastomas, the most frequent embryonal brain tumor in the pediatric population. Fiser noted that the EPIC arrays for Illumina sequencing often require at least eight samples and, therefore, are not suitable in emergency situations. The team tried whole genome nanopore sequencing with ONT. The researchers compared forty-five medulloblastomas with this approach and obtained good concordance with Illumina EPIC arrays. Furthermore, they were able to multiplex six samples per MinION flow cell. The team then expanded to over 140 samples and found good concordance even at the subgrouping level. This confirmed that the methylation data from ONT is comparable to the current gold standard. The major advantage is that this approach was easy to implement with MinION devices.
